The purpose of this project was to investigate if an improvement of the time consumed to move goods in a specific grocery warehouse can be made by changing and adding factors to the current allocation model. After analyzing raw data collected from the warehouse and constructing new models, the factors of interest could be evaluated. Evaluation of the factors was made by simulations and statistical testing. Factors that were investigated were the turnover for the different articles and an ABC classification made on that factor. Changes in articles saturation for the aisles were also considered. The findings may give some useful insights of paths for further investigation and future improvement of the logistic system of the warehouse.

This dissertation focuses on developing analytical models for automated storage and retrieval system with multiple in-the-aisle pick positions (MIAPP-AS/RS). Specifically, our first contribution develops an expected travel time model for different pick positions and different physical configurations for a random storage policy. This contribution has been accepted for publication in IIE Transactions (Ramtin & Pazour, 2014) and was the featured article in the IE Magazine (Askin & Nussbaum, 2014). The second contribution addresses an important design question associated with MIAPP-AS/RS, which is the assignment of items to pick positions in an MIAPP-AS/RS. This contribution has been accepted for publication in IIE Transactions (Ramtin & Pazour, 2015). Finally, the third contribution is to develop travel time models and to determine the optimal SKUs to storage locations assignment under different storage assignment polies such as dedicated and class-based storage policies for MIAPP-AS/RS. An MIAPP-AS/RS is a case-level order-fulfillment technology that enables order picking via multiple pick positions (outputs) located in the aisle. We develop expected travel time models for different operating policies and different physical configurations. These models can be used to analyze MIAPP-AS/RS throughput performance during peak and non-peak hours. Moreover, closed-form approximations are derived for the case of an infinite number of pick positions, which enable us to derive the optimal shape configuration that minimizes expected travel times. We compare our expected travel time models with a simulation model of a discrete rack, and the results validate that our models provide good estimates. Finally, we conduct a numerical experiment to illustrate the trade-offs between performance of operating policies and design configurations. We find that MIAPP-AS/RS with a dual picking floor and input point is a robust configuration because a single command operating policy has comparable throughput performance to a dual command operating policy. As a second contribution, we study the impact of selecting different pick position assignments on system throughput, as well as system design trade-offs that occur when MIAPP-AS/RS is running under different operating policies and different demand profiles. We study the impact of product to pick position assignments on the expected throughput for different operating policies, demand profiles, and shape factors. We develop efficient algorithms of complexity O(nlog(n)) that provide the assignment that minimizes the expected travel time. Also, for different operating policies, shape configurations, and demand curves, we explore the structure of the optimal assignment of products to pick positions and quantify the difference between using a simple, practical assignment policy versus the optimal assignment. Finally, we derive closed-form analytical travel time models by approximating the optimal assignment's expected travel time using continuous demand curves and assuming an infinite number of pick positions in the aisle. We illustrate that these continuous models work well in estimating the travel time of a discrete rack and use them to find optimal design configurations. As the third and final contribution, we study the impact of dedicated and class-based storage policy on the performance of MIAPP-AS/RS. We develop mathematical optimization models to minimize the travel time of the crane by changing the assignment of the SKUs to pick positions and storage locations simultaneously. We develop a more tractable solution approach by applying a Benders decomposition approach, as well as an accelerated procedure for the Benders algorithm. We observe high degeneracy for the optimal solution when we use chebyshev metric to calculate the distances. As the result of this degeneracy, we realize that the assignment of SKUs to pick positions does not impact the optimal solution. We also develop closed-form travel time models for MIAPP-AS/RS under a class-based storage policy.

Submitted by JOSIANE SANTOS DE OLIVEIRA (josianeso) on 2017-06-20T11:56:48Z No. of bitstreams: 1 Carlos Eduardo Martins Machado_.pdf: 1371444 bytes, checksum: c6dbd9395f1c37af92e5c1fe24590ee7 (MD5) / Made available in DSpace on 2017-06-20T11:56:48Z (GMT). No. of bitstreams: 1 Carlos Eduardo Martins Machado_.pdf: 1371444 bytes, checksum: c6dbd9395f1c37af92e5c1fe24590ee7 (MD5) Previous issue date: 2017-05-26 / Nenhuma / A busca por novos mercados, novas tecnologias e a redução dos custos são necessárias para que as empresas consigam sobreviver no cenário de globalização onde as mudanças são constantes. Este trabalho teve como objetivo identificar os critérios utilizados por empresas transnacionais para decisão de alocação de produtos em suas fábricas previamente instaladas, visando a maximização do desempenho econômico-financeiro global. Foi realizado um Estudo de Caso único com objetos incorporados em uma empresa transnacional da indústria de roadbuilding. Foram pesquisadas 3 fábricas localizadas no Brasil, China e Índia, além do headquarter localizado na Suíça. Verificou-se que a gestão da empresa é feita através de decisões seletivas que visam obter vantagens competitivas globais em termos de custos e receitas. Pode-se afirmar que os critérios relacionados à perspectiva econômica, são os mais relevantes no processo decisório de alocação de produtos em suas subsidiárias. A redução dos custos diretos e indiretos relacionados aos produtos, a redução dos custos logísticos e a sobreposição de barreiras tarifárias foram os critérios mais relevantes identificados durante esta pesquisa. Estes critérios indicam a necessidade da empresa de avaliar o custo final ao cliente, para o cálculo de suas margens. Portanto, indica-se o que foi chamado de “landed cost” como um elemento importante na decisão de alocação, sendo este influenciado pelos três principais critérios citados anteriormente. O acesso a manufaturas de baixo custo, acesso a habilidades e conhecimento e acesso ao mercado são critérios relevantes ao caso de decisão de alocação de produtos em plantas previamente existentes. Apresenta-se um modelo geral que leva em conta os inter-relacionamentos de uma Matriz Global de Produtos vs. Geografia vs. Planta de produção. A maximização da performance econômico financeira global se dá através de decisões equilibradas nas perspectivas - manufatura, econômica e ambiental, alinhada com o objetivo estratégico da alocação do produto. / The search for new markets, new technologies and cost reductions are necessary for companies to survive in a globalized scenario where changes are constant. The target of this research was to identify the criteria used by transnational companies to decide about the product allocation in their factories, aiming to maximize global economic and financial performance. A Single Case Study was conducted with multiple unit of analysis (embedded design) in a transnational company in the roadbuilding industry. Three factories located in Brazil, China and India were surveyed, as well as the headquarter located in Switzerland. It was verified that the product allocation is made through selective decisions, which aims to obtain global competitive advantages in terms of costs and revenues. It can be affirmed that the criteria related to the economic perspective are the most relevant in the decision-making process of product allocation in its subsidiaries. The reduction of direct and indirect costs related to products, reduction of logistics costs and the overlapping of tariff barriers were the most relevant criteria identified during this research. These criteria indicate the needs of the company to evaluate the final cost for end-customer in order to calculate its margins. Therefore, what is called “landed cost” is indicated as an important element in the allocation decision, which is influenced by the three main criteria mentioned above. Access to low-cost manufacturing, access to skills and knowledge, and access to the market are relevant criteria to the product allocation decision in previously existing plants. It is presented a general model, which takes into account the relationships of a Global Product Matrix vs. Geography vs. Production plant. The maximization of global financial-economic performance is made through balanced decisions in the manufacturing, economic and environmental perspectives, aligned with the strategic objective of product allocation.

In this study we have conducted research on how to optimize inventory management within logistics. The focus in this study is to examine the picking rounds, the reason for this is because it is the most time consuming and expensive part within a warehouse. Is it possible to minimize the handling time to create efficient picking rounds? As a part of the research project AI has been investigated as well. If it is possible with help of AI, create a streamlining of current warehouse logistics. The purpose of this report is to investigate how to minimize the distance in picking rounds for efficient warehouse management. To be able to fulfil the purpose of the report research questions where conducted. The methodology that was chosen at first was traditional data collection. With the help of other studies conducted in this area we started to collect information. To be able to compare this information to the chosen company Care of Carl a case study was performed. A case study on the current situation at Care of Carl, and what the current optimization is based on. With the help of these two methods a result emerged. The result that was conducted by this study is that placement and categorization of products as well as route planning has a significant role when streamlining the picking process and minimizing the picking process. To store items in a warehouse the most suitable option is to use a free item placement, or storage out of sale frequency. But important to acknowledge is that it requires support systems to make this storage possible. When categorizing articles, it is crucial to combine this with a suiting picking method. In the case study, combining ABC categorization with zone picking was a possible solution. In general, it might be a good idea to invest in AI to use the picking position principle. With AI it is possible to analyse more complex data such as customer patterns and if this implementation succeeds it can lead to great advantages within a warehouse and the picking processes. The traveling distance constitutes most of the total picking time, it is important to have a route method that works with how you have chosen to place the items. This study shows that the optimal routing method is the one to use. This study showed that there are a lot of different strategies and methods on the current market. According to the case study Care of Carl can make big savings by changing strategies and methods. The reason why is because they have been reactive when investing in IT support systems. But in general, if a company wants to meet the current increasing requirements according to the globalization and the continuous changes within logistics operations, AI is the next step. The methods that are currently used are not sufficient, with the help of AI there is room for improvements within product allocation and route planning. / I denna studie har det undersökts hur man kan optimera lagerhanteringen inom logistik. Fokus har varit att undersöka plockrundorna, då det är den mest tidskrävande och kostsamma delen inom ett lager. Är det möjligt att minimera hanteringstiden och därmed effektivisera plockrundorna? Studien har även varit en del av ett forskningsprojekt där man har undersökt om det med hjälp av AI är möjligt att skapa en effektivisering av lagerhantering. Syftet med denna rapport är att undersöka hur man minimerar avståndet i plockrundorna för att effektivisera lagerhanteringen. För att kunna uppfylla syftet med rapporten utformades det forskningsfrågor kopplat till syftet. Traditionell datainsamling var den metod som användes för att komma i gång med studien. Den teoretiska referensramen som skapades i denna rapport var utifrån andra studier som genomförts inom detta område, men även utifrån att kunna besvara de forskningsfrågor som skapats. Det genomfördes även en fallstudie på företaget Care of Carl, med en nulägesbeskrivning samt en förklaring gällande hur deras nuvarande optimering tagits fram. För att kunna besvara syftet med rapporten och forskningsfrågorna jämfördes den teoretiska referensramen med den fallstudien som genomförts i samband med denna studie. Resultatet som framkom under studien var att placering och kategorisering av produkter såväl som ruttplanering har en avgörande roll gällande effektivisering av plockprocessen i ett lager. Gällande inlagringsmetod är det lämpligast att använda sig av flytande artikelplacering alternativt lagring utifrån försäljningsfrekvens. Vad som är viktigt att nämna är att båda metoder kräver ett stödsystem för att kunna implementeras. Gällande kategorisering av artiklar är det viktigt att kombinera detta med en passande plockmetod. I fallstudien var en möjlig lösning att kombinera ABC-kategorisering med zonplockning. Generellt sätt är AI en framtida värd investering då man kan använda sig av plockpositionsprincipen. AI möjliggör analysering av mer komplexa data som kundmönster och om denna implementering lyckas kan det leda till stora fördelar inom ett lager och för plockprocessen. Det är även viktigt att ha en ruttmetod som fungerar ihop med den placeringsmetod man använt sig av, då gångtiden och gångavståndet är det som utgör det mesta av den totala plocktiden. Denna studie visar att den optimala ruttmetoden är den som bör användas, och detta kräver en investering i ett stödsystem. Denna studie visade att det för tillfället finns många olika strategier och metoder på marknaden idag. Enligt fallstudien kan Care of Carl göra stora besparingar bara genom att ändra sina strategier och metoder. Orsaken är att de har varit reaktiva vid investeringav IT-stödsystem. Generellt sätt, om ett företag vill uppfylla de ökande kraven som finns till följd av globaliseringen och de kontinuerliga förändringarna inom logistikverksamheten, är AI nästa steg att ta. Metoderna som för närvarande används är inte tillräckliga och med hjälp av AI finns det möjlighet för förbättringar inom produktallokering och ruttplanering.

This case study is investigating the challenge of managing product scarcity in the supply chain by allocating scarce products to customers while maintaining customer relationships in the Fast-Moving Consumer Goods (FMCG) sector. Previous solutions on how to manage the situation of allocating scarce products have been inadequate, highly manual, and timeconsumingwork for all stakeholders. The purpose of this study is to investigate the challenges of achieving an effective supply chain during scarce and uncertain conditions and delivering a developed Business Continuity Plan (BCP) to cope with these challenges, by answering the research question: How should a Business Continuity Plan be set up to handle the allocation of scarce products to remain relevant and keep customers satisfied? A thematic literature review combined with unstructured and semi-structured interviews provided data showing that product scarcity is extra critical in categories such as electronics with highly seasonal demand combined with longer lead times and higher profitability. The most important factors to handle product allocation toward customer is the need for a more structured process, a high level of automation, and increased visibility. To make sure that the new process is anchored in the organization, raising awareness, having extensive onboarding sessions, and having good documentation on the processes were highlighted. The developed solution is a highly automated process in KNIME generating the allocation and visualizing it combined with other data in a Power BI report, increasing visibility and enhancing communication, answering the needs explained in the interviews which are highly aligned with the literature. / Denna studie undersöker utmaningen med att hantera produktbrist i leveranskedjan genom att allokera knappa produkter till kunder samtidigt som kundrelationer upprätthålls inom sektorn för snabbrörliga konsumentvaror (FMCG). Tidigare lösningar på hur man hanterar situationen med att tilldela knappa produkter har varit otillräckligt, manuellt och tidskrävande arbete för alla inblandade. Syftet med denna studie är att undersöka utmaningarna med att uppnå en effektiv försörjningskedja under knappa och osäkra förhållanden och att leverera en utvecklad Business Continuity Plan (BCP) för att klara av dessa utmaningar, genom att svara på forskningsfrågan: Hur ska en Business Continuity Plan konstrueras för att hantera tilldelning av knappa produkter för att förbli relevanta och hålla kunderna nöjda? En tematisk litteraturgenomgång i kombination med ostrukturerade och semistrukturerade intervjuer gav data som visade att produktbrist är extra kritiskt för kategorier inom elektronik med säsongsbetonad efterfrågan i kombination med längre ledtider och högre lönsamhet. De viktigaste faktorerna för att hantera produktallokering mot kund är behovet av en mer strukturerad process, hög grad av automatisering och ökad transparens. För att säkerställa att den nya processen förankras i organisationen lyftes kunskap, omfattande onboarding-sessioner och bra dokumentation om processerna fram. Den utvecklade lösningen är en automatiserad process i KNIME som genererar allokeringen och visualiserar den i kombination med annan data i en Power BI-rapport, vilket ökar synligheten och förbättrar kommunikationen, och svarar på de behov som lyfts fram i intervjuerna, i linje med den presenterade litteraturen.

Lagerhantering anses vara en betydelsefull aktivitet som företag måste ta sig an, då förbättrade metoder kan ha direkt påverkan på resultatet. Produktallokering inom lagret är en av de vitala delarna inom lagerplanering. Detta arbete är en fallstudie där fallet är ett lager hos ett producerande företag med det övergripande syftet att undersöka hur en optimeringsmodell för lagerhantering kan hjälpa organisationer med dess lagerhantering genom att studera rådande processer och utifrån det skapa en optimeringsmodell för att beräkna optimal lagerallokering. Studien har utförts genom en mixad metod, uppdelad i en kvantitativ- och en kvalitativ fas, då forskning hävdar att nyckeln till framgång för effektiv lagerhantering är att matcha policyn med lagrets layout och de ordertyper som lagret hanterar, för att anpassa optimeringsmodellen utifrån det givna fallets policy, lagerlayout och ordertyper. I den första fasen genomfördes en processkartläggning för att bygga upp för optimeringens utformning. Den andra fasen gick ut om att matematiskt beskriva samma problemuppställning och modellera problemet på insamlad kvantitativ data. I processkartläggningsfasen identifierades de två differentieringskriterierna omsättning i enheten sålda paket och förädling i enheten produktionskostnad. Målet med utförd klassificering var att ta fram vilka unika produkter som bör behandlas differentierat och erhålla tilldelad lagring, vilket var en förbättringsmöjlighet funnen i nulägesbeskrivningen. Totalt togs det fram 93 produkter att placera ut på någon av de 220 lagringsplatserna med målet att minimera körsträckan med truck. Jämförelse av målfunktionsvärden mellan optimaluppsättning enlig modell och dagens uppsättning, påvisade en teoretisk besparing av antal körda meter på 35 %. Studiens resultat har påvisat att matematisk modellering och linjärprogrammering går att använda för lagerallokeringsproblem för att optimera produktallokeringen, där den matematiskmodellen baserats på fallspecifika egenskaper. Modellens resultat förväntas underlätta ett antal av de problem som noterats vid kartläggning av lagerhanteringsprocesserna. / Inventory management is considered to be an important activity that companies need to accomplish as improved methods directly can affect the final result. Product allocation within the warehouse is one of the vital parts of inventory management. This report is a case study where a warehouse of a manufacturing company is considered the case. The overall purpose is to investigate how a optimization model can help organizations with their inventory management, by first studying existing processes and based on that create the model to calculate optimal inventory allocation. The study has been conducted through a mixed method, divided into a quantitative and qualitative phase, as research claims that the key to the success of efficient inventory management is to match the layout of the warehouse case policy, stock layout and order types in order to customize the optimization model. In the first phase, a process mapping was conducted to build the optimization design. The second phase was based on mathematically describing the same problem setup, and then to model the problem with collected data. In the process-mapping phase, two differentiation criteria were identified as units sold packages and the unit production cost. The purpose of the classification was to identify which unique products that should be treated differently and receive assigned storage, which was an improvement opportunity found in the case description. A total of 93 products were placed on one of the 220 storage sites, with the aim of minimizing the mileage by truck. Comparison of target function values ​​between the optimal set-up model and the current set, demonstrated a theoretical saving of the number of driven meters of 35%. The study results have shown that mathematical modeling and linear programming can be used for inventory allocation problems to optimize product allocation, where the mathematical model is based on case-specific properties. The model's results are expected to facilitate a number of the problems noted in mapping inventory management processes.